Nuke,
You are absolutely correct that we are not familiar with the actual process of coding and distribution throughout the treatment cycles, nor do we have knowledge of exactly what happened, because we were not given that information; however I go back to my earlier observation that if you mess something up, you do not obtain good results. Look at the MOS results. If CSM was "randomly assigning drugs in a continuously varying manner" or an "intern running the coding and continuously screwing it up" would it be likely to see that kind of separation, or would you expect all three arms to come in about equal? Again, reminds me of the OJ Simpson trial. OJ's DNA matched the blood evidence from the crime scene - the defense argued that the police botched the handling of the blood evidence. If they botched it, then how would OJ's DNA match it, with a 10 billion to 1 odds that it could have come from someone else? If they botched it, then nobody's DNA should have matched it.